Adaptive Transmit Queue Length for Packet Loss Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication networks face challenges in reducing data packet loss, particularly in wireless WAN interfaces, due to the lack of cross-layer awareness and the need for network proxy or management node components, which limits simultaneous use of multiple interfaces for transmit queue management.
Innovation Solution
A cross-layer system that adapts the transmit queue length of multiple active interfaces based on network conditions, using a network driver module to estimate effective bandwidth ratios and adjust queue lengths proportionally, without requiring modifications to the physical or data-link layers or any intermediary nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If transmit queue management is implemented using existing systems, then packet loss can be reduced, but the system requires network proxy or management node components and cannot simultaneously use multiple interfaces
Solution Approach 1:
The system enables self-service by implementing transmit queue management directly at the network interface card (NIC) level without requiring external network proxy or management node components. Each NIC independently manages its own transmit queue based on bandwidth aggregation estimates, eliminating the need for intermediary control infrastructure while maintaining packet loss reduction capabilities
Solution Approach 2:
The system segments the transmit queue management function across multiple independent network interfaces, allowing each interface to be managed separately with its own adaptive queue length. This segmentation enables simultaneous use of multiple interfaces for packet transmission while maintaining individual queue control, resolving the contradiction between packet loss reduction and system complexity
2Reliability
If adaptive transmit queue length is used, then packet loss is reduced in poor network conditions, but the system requires cross-layer awareness and bandwidth aggregation mechanisms
Solution Approach 1:
The system implements dynamic adaptability by continuously adjusting transmit queue lengths based on real-time bandwidth aggregation estimates from multiple network interfaces. The queue length adapts dynamically to changing network conditions without requiring complex cross-layer protocols, achieving packet loss reduction through adaptive response to measured bandwidth availability
Solution Approach 2:
The system achieves universality by implementing a multi-functional bandwidth aggregation mechanism that simultaneously performs network condition monitoring, queue length adaptation, and packet transmission across multiple interfaces. This universal approach handles various network conditions and interface types without requiring specialized cross-layer awareness mechanisms
3Productivity
If multiple active interfaces are used for bandwidth aggregation, then transmission capacity increases, but queue management becomes more complex without cross-layer awareness
Solution Approach 1:
Each network interface independently performs self-service queue management by monitoring its own bandwidth contribution to the aggregation and adjusting its transmit queue length accordingly. This distributed self-service approach simplifies multi-interface queue management by eliminating the need for centralized coordination or complex cross-layer awareness while maintaining optimal queue lengths across all interfaces
Data Source
Figure 1
Figure 2
AI summary
The present invention provides a system and method for reduction of data packet loss for the multiple network interfaces. Particularly, the invention provides a cross layer system for reduction of data packet loss based on dynamic analysis of network conditions. Further, the invention provides a system and method of estimation of network condition and adapting the transmit queue of the multiple interfaces according to channel condition / available bandwidth of the associated network.